6 Analyzes and Results
6.4 Test of the Second Research Model; Determinants of CEO compensation
performance that affect and determine CEO compensation. In this subchapter, we will test our hypotheses for our second research model, where we want to examine determinants of CEO compensation, where we divide CEO compensation in both variable CEO compensation and fixed salary, by using multiple regression analyzes.
We start by presenting the tables of the findings where we test the relationship between the different independent variables, as firm size measured by market value and revenue, firm risk measured by beta, CEO's direct ownership, CEO's age, CEO's tenure, board size and board gender and the dependent variable CEO compensation, measured as variable CEO
compensation. Further, we will present the table of the findings from the multiple regression where we test for the independent variables’ effect on the dependent variable, measured as CEO's fixed salary. We test all of the independent variables at the same time and examine which of these variables that have the biggest effect on variable CEO compensation and CEO's fixed salary. We will explain the tables below before we discuss each of the
hypotheses in more detail, as the hypothesis are not divided for fixed salary and variable CEO compensation.
Table 6.17.1 – Regression analysis for research model two (Variable)
LN_Var is the dependent variable and the natural logarithm of variable CEO compensation, LN_MV is the natural logarithm of market value, LN_Rev is the natural logarithm of revenue, LN_Beta is the natural logarithm of beta, LN_OS is the natural logarithm of CEO’ direct percentage ownership in the firm, LN_Age is the natural logarithm of the CEO’s age, LN_Ten is the natural logarithm of the CEO’s tenure measured in years, LN_BS is the natural logarithm of board size, which is the total number of directors in the board, LN_BG is the natural logarithm of the percentage of female directors in the board. The observations in this table are for all of the four years, 2010-2013.
127 Minu Singh and Cigdem Yavuz
Table 6.17.1 shows the results of the multiple regression analysis for our dependent variable, variable CEO compensation, and different independent variables which can be determinants of variable CEO compensation. These variables are transformed, to make sure that the error term meets the requirements of normal distribution, which we discuss under regression assumption 8 in subchapter 6.5.8.
The results from the table show that the different independent variables have an explanatory power of 17.3 % on the dependent variable, variable CEO compensation. This means that there are other variables that explain more of variable CEO compensation, than the variables we have chosen to examine. The total number of observations has additionally decreased down to 158 in this regression analysis as there are extreme values that are removed, and all negative numbers and numbers below 0 cannot be transformed, and have become missing values. Even though the observation number has decreased, and can affect our results, we consider that it would be more critical if we did not remove the extreme values or transformed our variables, based on the previous discussions. Further, we see from the table that there only are two independent variables which are significant, and we will discuss these in detail when we present our hypotheses. We will first present the table below that shows the independent variables’ effects on CEO’s fixed salary.
Table 6.17.2 – Regression analysis for research model two (Fixed)
LN_Fix is the dependent variable and the natural logarithm of the CEO’s fixed salary, LN_MV is the natural logarithm of market value, LN_Rev is the natural logarithm of revenue, LN_Beta is the natural logarithm of beta, LN_OS is the natural logarithm of CEO’ direct percentage ownership in the firm, LN_Age is the natural logarithm of the CEO’s age, LN_Ten is the natural logarithm of the CEO’s tenure measured in years, LN_BS is the natural logarithm of board size, which is the total number of directors in the board, LN_BG is the natural logarithm of the percentage of female directors in the board. The observations in this table are for all of the four years, 2010-2013.
128 Minu Singh and Cigdem Yavuz
Table 6.17.2 shows the results of the multiple regression analysis for our dependent variable, measured as CEO fixed salary. Findings from the table show that the different independent variables have a high explanatory power of 55.3 % on the dependent variable, CEO fixed salary. This means that the variables we have chosen explain half of the variation in fixed salary, while the rest is explained by other variables. The total number of observations has gone up by two more observations from 158 to 160 in this regression analysis. Further, we see from the table that there are now three independent variables which have a significant effect on the dependent variable, and we will discuss these in more in detail when we now present our hypotheses. We will discuss the findings from both the tables above, with the independent variables' effect on variable CEO compensation and fixed salary under each hypothesis.
HA: Firm size has a positive effect on CEO compensation
In the tables above we have tested how the independent variable, firm size, affects the dependent variable, CEO compensation, measured by both variable CEO compensation and fixed salary. Firm size is still measured by market value and revenue, and our findings show that both of these measures have a positive effect on variable CEO compensation and fixed salary. As we see, market value is positive and significant at the 0.01 level, while revenue only is positive and not significant. This indicates that firm size, measured my market value, has a positive effect on CEO compensation.
From the theoretical aspects of the managerial power theory and corporate governance, firm size has a direct effect on CEO compensation. Large firms tend to have multiple owners, with spread ownerships, where the CEOs can dominate and gain more control over the board of directors, and eventually increase their own compensation. Additionally, large firms have more resources and capable of giving higher compensation to the CEOs (Berle & Mean, 1933;
Gomez-Mejia et al., 1987). Additionally, CEOs that work in larger firms tend to have more responsibilities, so this can also explain why CEOs have higher fixed salary in large firms.
This is also consistent with the empire building theory, where CEOs try to increase the firm size in order to get higher compensations. Hence, our results are consistent with the theories, and firm size has a positive effect on CEO compensation in our study.
Conclusion: As we find significant results of firm size’s effect on CEO compensation, we keep hypothesis A.
We will further discuss the findings of how firm risk affects CEO compensation.
129 Minu Singh and Cigdem Yavuz
HB: Firm risk has a positive effect on CEO compensation
We test firm risk’s effect on variable CEO compensation and fixed salary. The findings show that there is a positive effect between firm risk and variable CEO compensation, but this result is not significant. We can hence not conclude that there is a significant effect of firm risk on variable compensation. However, the results are both positive and significant for fixed salary on the 0.05 level, which indicates that CEOs who work in firms with larger risk, get higher fixed salaries.
From the classical principal-agent theory we know that in high-risk firms the shareholders have to give more incentives to the CEO in order to relive the CEOs for risk. With proper incentives as higher compensation, the shareholders can try to reduce the risk. We expected that this could affect the variable part of the CEO compensation more, as these are more related to incentives. However, the shareholders can also release the CEOs from risk by giving them higher fixed salaries, which will not be adjusted for many factors, and gives them a feeling of security. Further, CEOs in high-risk firms will have more responsibilities and meet more challenges, which indicates higher fixed salaries. Since our findings show
consistence with the theory, we see that firm risk does affect CEO compensation in practice.
Conclusion: As we find significant results of firm risk's effect on CEO compensation, we keep hypothesis B.
We will now discuss hypothesis C in more detail.
HC: CEO's direct ownership has a positive effect on CEO compensation
The tables above also show the regression analyzes for how CEO's direct ownership affects variable CEO compensation and fixed salary. The findings are positive, but not significant, for variable compensation, and negative and significant for fixed salary on the 0.01 level.
Hence, CEO’s direct ownership has a negative effect on CEO compensation, when we measure CEO compensation by fixed salary.
We get interesting findings, as on the basis of the managerial power theory and aspects of corporate governance we expected to find that CEO’s direct ownership would have a positive effect on CEO compensation, measured by variable CEO compensation and fixed salary. The reason for this is that when the CEOs own many firm shares, they will have more influence on the director elections and be more able to determine and negotiate their own compensations.
However, our findings are inconsistent with the theory, and show that CEOs with higher
130 Minu Singh and Cigdem Yavuz
direct ownerships get lower fixed salaries. The reason for this can be that the CEO is more willing to get a lower fixed salary when he has direct ownerships in the firm in order to not threaten the firm’s economy, as this will affect himself as well. Our findings are not consistent with the theory, and we can hence not conclude that CEO’s direct ownership has a positive effect on CEO compensation.
Conclusion: As we do not find both positive and significant results of the CEO's direct ownership on CEO compensation, we reject hypothesis C.
We will further test how CEO’s age and tenure affect CEO compensation. Our main intention was also to test how CEO’s gender affect CEO compensation, but as mentioned we have very few observations of female CEOs in firms listed on the Oslo Stock Exchange. Hence, we are not able to test hypothesis D3.
HD1: CEO’ age has a positive effect on CEO compensation
We have tested the effect of CEO's age on CEO compensation, for both variable CEO compensation and fixed salary. From our regression analysis we find that there is a positive but not a significant relationship between the CEO's age and variable CEO compensation.
Further, the results indicate that the CEO’s age decreases the fixed salary, but these findings are neither significant. We expected to find a positive and significant effect of CEO age on CEO compensation, as this is consistent with the human capital theory. When the CEO's get older they get more experienced and thereby can take more responsibilities which will lead to higher compensation. Higher responsibilities should also give higher fixed salary, as the CEOs meet more challenges. Hence, we are surprised that we do not find any positive and significant results in our study between CEO age and CEO compensation. We cannot think of an economical rationality behind these results, and we can hence not keep our hypothesis.
Conclusion: As we do not find a positive and significant effect of CEO’s age on CEO compensation, we reject hypothesis D2.
131 Minu Singh and Cigdem Yavuz
HD2: CEO’s tenure has a positive effect on CEO compensation
In the tables above, we can also see how CEO’s tenure affects the CEO compensation. The findings show that there is a negative and no significant effect between CEO tenure and CEO compensation, for both variable CEO compensation and fixed salary. From a human capital view and from the managerial power theory, a longer CEO tenure will increase the CEO's compensation, since CEOs who sit in a position over a longer time will gain more knowledge and power, and can thereby influence the board of directors to increase their compensation.
Additionally, it is common to think that CEOs who sits in the same position over time, get higher fixed salary through the years as they get more experienced, and as they get more knowledge of how the firm should be managed (Blaug, 1976; Boyd, 1994; Hill & Phan, 1991;
Schultz, 1961; Zajac & Westphal, 1996).
Our results on the other hand, indicates that CEOs get lower compensation as their tenure increases, which is inconsistent with the theory. These findings are however not significant, similar to the findings of Randøy and Skalpe (2007). We hoped to achieve different results that indicated a positive and significant relationship between CEO tenure and CEO
compensation, but we find no significant effect between these two variables that can be explained by economical rationality. Hence we have to reject this hypothesis.
Conclusion: As we do not find both a positive and significant effect of CEO’s tenure on CEO compensation, we reject hypothesis d2.
We will further test how board size and number of female directors in the board affect CEO compensation. As in the first research model, we cannot test how publicly and privately owned firms affect CEO compensation, as there are few observations of publicly owned firms, firms where the Norwegian state and government owns shares directly in the firm.
HF1: The size of the board of directors will have a positive effect on CEO compensation We have also tested how the size of the board of directors can affect CEO compensation. In the tables above we find a negative and not significant effect of board size on variable CEO compensation, and a positive but not significant effect of board size on fixed salary. This indicates that the size of the board of directors has no significant effect on CEO compensation in our study.
From the managerial power theory and corporate governance aspects in Norway, the board of directors is responsible of determining the compensation of the CEOs. If there are many
132 Minu Singh and Cigdem Yavuz
directors in the board, the board will be easier to influence and the CEO can gain more power, and thereby determine their own compensation on a bigger scale (Jensen, 1993). The results we get are similar to the findings of Randøy and Skalpe (2007), and are inconsistent with the theory which indicates that the board size does not affect CEO compensation in practice.
Hence, we unfortunately have to reject this hypothesis as the size of the board of directors does not have any effect on CEO compensation.
Conclusion: As we do not find both a positive and significant effect of board size on CEO compensation, we reject hypothesis F1.
HF2: The number of female directors in the board will have a positive or a negative effect on CEO compensation
In the tables above we see that number of female directors in the board has a negative and significant effect on the 0.05 level on CEO compensation, measured by variable CEO
compensation. The findings show however no significant relationship between the number of female directors and fixed salary. This means that our results show what we expected to one point, that a greater portion of female directors in the board decreases CEO compensation.
It is interesting to find a negative effect, which in accordance to the theory means that when number of female directors increases, the CEO compensation decreases. The reason behind this, can for instance be that the female directors want equality and are not easily influenced by the CEO as the majority of the CEOs in firms listed on the Oslo Stock Exchange are males.
In our first research model we found that the number of female directors weakens the pay-for-performance relationship, and that CEOs get their variable CEO compensation independent of firm performance. However, even if they receive compensation independent of firm
performance, they still get lower variable CEO compensation than in firms where there are less female directors. This indicates female directors have more control and moderate the CEO compensation. Our findings partly support our hypothesis, where the number of female directors in the board has a negative and a significant effect on CEO compensation.
Conclusion: As we find both a negative and significant effect of number of female directors on CEO compensation, we keep hypothesis F2.
In the below table we present the results from our second research model.
133 Minu Singh and Cigdem Yavuz
6.4.1 Summary of Results for the Second Research Model
Hypotheses Results
HA: Firm size has a positive effect on CEO compensation
Confirmed
Market value is positively significant with variable CEO compensation and
fixed salary HB: Firm risk has a positive effect on CEO
compensation
Confirmed
Firm risk is positively significant with fixed salary
HC: CEO's direct ownership has a positive effect on CEO compensation
Rejected
CEO’s direct ownership is negatively significant with fixed salary HD1: CEO’ age has a positive effect on CEO
compensation
Rejected No significant results
HD2: CEO’s tenure has a positive effect on CEO compensation
Rejected No significant results
HD3: CEO compensation is higher for male CEOs than female CEOs
Cannot be tested
HE: CEO compensation is lower in publicly owned firms than in privately owned firms
Cannot be tested
HF1: The size of the board of directors will have a positive effect on CEO compensation
Rejected No significant results
HF2: The number of female directors in the board will have a positive or a negative effect on CEO
compensation
Confirmed
The number of female directors is negatively significant with variable
CEO compensation
Table 6.18 – Results for the second research model
134 Minu Singh and Cigdem Yavuz